In June 2007, investigators visited the veterinary clinic and the household of the index cat and conducted a contact investigation and carrier study. Interviews of 2 household members and 8 veterinary staff members indicated no recent respiratory illness, skin infection, or risk factors for diphtheria (e.g., travel to countries to which diphtheria is endemic or contact with known case-patients). Half of these 10 contacts had received diphtheria vaccination within the previous 5 years. Cultures of oropharyngeal swab samples obtained from each person were negative, including cystine tellurite blood agar, which is selective for C. diphtheriae. Household members also were interviewed about medical history of a convenience sample of household animals (4 cats, including the index cat; 2 dogs; and 1 horse). Each animal was briefly examined, and oropharyngeal, otic, or ocular swab samples were collected. Otitis was observed in all 4 cats and 1 dog. The horse reportedly had had an eye infection ≈5 years earlier. No other abnormal findings were noted. Animal specimens yielded 3 additional isolates of C. diphtheriae: 1 from each ear of the index cat and 1 from the left ear of a 2-year-old domestic medium-hair cat. Both cats had been born on the premises and had remained with the same household since birth.

Feline C. diphtheriae and reference isolates used are described in the Table. Tinsdale agar plate growth (Remel, Lenexa, KS, USA) gave rise to black colonies with a brown halo, typical of cysteinase-producing C. diphtheriae, C. ulcerans, or C. pseudotuberculosis. After 24 hours on blood agar, 1–2-mm grey-white or opaque, rounded, convex colonies with no hemolysis were observed. Microscopically, the bacteria were gram-positive, club-shaped rods, 1 µm in diameter, arranged singly or at angles. Biochemical profiles to determine species and biotype were done by using an API Coryne strip (bioMérieux, Durham, NC, USA, and St-Laurent, Quebec, Canada). Query of API Coryne code 0010304 obtained for all isolates by APIWEB (https://apiweb.biomerieux.com) indicated a decreased level of confidence of C. diphtheriae biotype mitis or belfanti (89.5%) because of a maltose-negative result. Isolates were further characterized morphologically and biochemically by using tube substrates (2) and were identified by using a standard taxonomic scheme (3). Feline isolates were biochemically identical with each other and phenotypically consistent with C. diphtheriae biotype belfanti, except for the lack of maltose fermentation, which was considered an unusual finding (3).

Antimicrobial drug susceptibility testing was performed according to the Clinical and Laboratory Standards Institute’s recommended methods and interpretative criteria (4). All 4 feline isolates were sensitive to ampicillin, cefepime, cefotaxime, ceftriaxone, cefuroxime, chloramphenicol, ciprofloxacin, clindamycin, daptomycin, erythromycin, ertapenam, gatifloxacin, gentamicin, levofloxacin, linezolid, meropenem, moxifloxacin, penicillin, quinupristin/dalfopristin, rifampin, telithromycin, tetracycline, tigecycline, trimethoprim/sulfamethoxazole, and vancomycin. Cellular fatty acid composition analysis was performed as described (5) by using the Sherlock system (MIDI, Inc., Newark, DE, USA), except that version 4.5 of the operating software was used. The cellular fatty acid composition profiles were consistent for C. diphtheriae, C. ulcerans, or C. pseudotuberculosis, including a substantial proportion (28%–30% of total) of C16:1ω7c (5). All feline isolates produced 7–15 meq/L of propionic acid among fermentation products, a feature associated with C. diphtheriae (2).

Results from use of the modified Elek test (6) indicated that all feline isolates were negative for production of diphtheria toxin; however, an atypical precipitation was observed after 36 h of incubation. Lack of toxin expression was corroborated by negative Vero cell assay results (7) and confirmed by using Western blot. Real-time PCR selective for the C. diphtheriae and C. ulcerans toxin gene (tox) (8) was positive for all feline isolates. However, real-time PCR for A and B subunits of tox (9) amplified subunit A but not subunit B. Sequence analysis of the tox gene was performed as previously outlined (10) and compared with a reference tox gene, GenBank accession no. K01722. The 4 feline tox sequences were identical to each other but contained multiple nucleotide substitutions and deletions compared with the reference gene. By NCBI BLAST search (http://blast.ncbi.nlm.nih.gov/Blast.cgi), the feline tox had higher sequence identity (97%–98%) to the tox sequences of C. ulcerans, compared with those from C. diphtheriae (94%–95%). A deletion at nt 55, coupled with a cytosine-to-thymine substitution at nt 74, prematurely terminated the peptide at aa 25.

Figure 1

Figure 1. Neighbor-joining phylogenetic tree based on 16S rRNA gene sequence analysis of Corynebacterium diphtheriae isolates, including 4 feline isolates from West Virginia, 2008 (ATCC BAA-1774, CD 448, CD 449, CD 450). The...

Figure 2

Figure 2. Jukes-Cantor–derived phylogenetic tree based on sequence analysis of a selected region of the rpoB gene of Corynebacterium isolates, including 2 feline isolates from West Virginia, 2008 (ATCC BAA-1774, CD 450). Feline...

Species characterization was corroborated by using 16S rRNA (11) and partial rpoB (12) gene sequencing. By 16S rRNA gene sequence analysis, the feline strains had 100% identity with each other and >99.1% identity with various reference sequences for C. diphtheriae biotype gravis and belfanti sequences, including NCTC 11397^T. Partial rpoB sequence analyses indicated 100% identity among the feline isolates and 97.7% identity with C. diphtheriae NCTC 11397^T. Neighbor-joining phylogenetic trees based on both 16S rRNA (Figure 1) and partial rpoB gene sequencing (Figure 2) positioned the feline isolate sequences within the C. diphtheriae clade but clearly distinguished them from the other C. diphtheriae isolates. Comprehensive molecular analyses to characterize differences between biotype belfanti strains, including these feline isolates, with other C. diphtheriae biotypes, are the subject of a separate publication (C.G. Dowson, pers. comm.).

We identified a potentially novel biotype of C. diphtheriae recovered from domestic cats in West Virginia but found no evidence of zoonotic transmission. Although rare, isolation of C. diphtheriae from animals has been reported, including C. diphtheriae biotype belfanti from a skin lesion of a cow (13) and toxigenic C. diphtheriae biotype gravis from a wound of a horse (14). C. ulcerans is a known animal pathogen, and zoonotic transmission of toxigenic C. ulcerans from companion animals has been reported, often associated with predisposing concurrent illnesses (15).

The feline strains isolated during this investigation differed phenotypically from previously described biotypes but were otherwise regarded as typical of C. diphtheriae. However, isolates were nontoxigenic and harbored a modified tox gene with sequence differences from Corynebacterium spp. capable of expressing diphtheria toxin. On the basis of published criteria (11), the feline strain might represent a novel subspecies of C. diphtheriae because it shares <98% sequence homology to the type strain within the rpoB gene. Potential for zoonotic transmission of this novel, cat-associated C. diphtheriae and associated public health implications are unknown. Additional studies are needed to further characterize these isolates and determine their appropriate taxonomy. Large-scale screening of domestic cat populations is recommended to determine the prevalence of C. diphtheriae and its pathogenic potential and to identify additional isolates for more formal description and classification.